Research map / Learning theory and algorithms / Prediction and decision losses
The Global Patch Collider
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add multi-scale learning to the single-scale GPC (ablation) | Global Patch Collider · Sintel Final, 25% recall | precision: 94.5% -> 98.1% | Helped |
| use a multi-tree forest rather than a single tree (ablation) | Global Patch Collider · Sintel Final, 25% recall | precision: 94.5% -> 93.6% | Helped |
| add multi-scale patches (ablation) | GPC sparse matching · Sintel sparse matching | endpoint error: 1.8796 -> 1.2809 | Helped |
| add motion-prior learning to multi-scale GPC (ablation) | GPC sparse matching · Sintel sparse matching | mean matches: 16813 -> 26131 | Helped |
| add motion-prior learning (ablation) | GPC sparse matching · IR stereo sparse matching | sparse endpoint error: 1.30 -> 1.00 | Helped |
| initialize PatchMatch with GPC correspondences instead of random initialization (ablation) | PatchMatch · Dense stereo, 1 iteration | error: 1.5940 -> 1.5863 | Helped |
| train task-specific GPC on Sintel and transfer it to wide-baseline stereo | GPC · Wide-baseline stereo, 1 frame | EPE: 0.12 -> 0.04 | Helped |
| train task-specific GPC on Sintel and transfer it to wide-baseline stereo | GPC · Wide-baseline stereo, 3 frames | inlier percentage: 71% -> 53% | Hurt |
| use motion priors with multi-scale GPC (ablation) | GPC sparse matching · Sintel sparse matching | endpoint error: 1.2809 -> 1.3626 | Hurt |
Closest papers by meaning
- GMSF: Global Matching Scene Flow (NeurIPS 2023 · similarity 0.86)
- RUBIK: A Structured Benchmark for Image Matching across Geometric Challenges (CVPR 2025 · similarity 0.84)
- World Tracing: Pixel-Aligned Geometry Beyond the Visible (NeurIPS 2026 · similarity 0.83)
- GLACE: Global Local Accelerated Coordinate Encoding (CVPR 2024 · similarity 0.83)
- Towards Complementary Keypoint Detection via Mixture of Detectors (NeurIPS 2026 · similarity 0.83)
- Zero-shot RGB-D Point Cloud Registration with Pre-trained Large Vision Model (CVPR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.